[大数据学习研究] 3. hadoop分布式环境搭建

1. Java安装与环境配置

Hadoop是基于Java的，所以首先需要安装配置好java环境。从官网下载JDK，我用的是1.8版本。 在Mac下可以在终端下使用scp命令远程拷贝到虚拟机linux中。

danieldu@daniels-MacBook-Pro-857  ~/Downloads  scp jdk-8u121-linux-x64.tar.gz root@hadoop100:/opt/software
root@hadoop100's password:
danieldu@daniels-MacBook-Pro-857  ~/Downloads 

其实我在Mac上装了一个神器-Forklift。 可以通过SFTP的方式连接到远程linux。然后在操作本地电脑一样，直接把文件拖过去就行了。而且好像配置文件的编辑，也可以不用在linux下用vi，直接在Mac下用sublime远程打开就可以编辑了 ：） 

然后在linux虚拟机中（ssh 登录上去）解压缩到/opt/modules目录下

[root@hadoop100 include]# tar -zxvf /opt/software/jdk-8u121-linux-x64.tar.gz -C /opt/modules/

然后需要设置一下环境变量， 打开 /etc/profile, 添加JAVA_HOME并设置PATH用vi打开也行，或者如果你也安装了类似forklift这样的可以远程编辑文件的工具那更方便。

vi /etc/profile 

按shift + G 跳到文件最后，按i切换到编辑模式，添加下面的内容，主要路径要搞对。

#JAVA_HOME
export JAVA_HOME=/opt/modules/jdk1.8.0_121
export PATH=$PATH:$JAVA_HOME/bin

按ESC ， 然后 ：wq存盘退出。

执行下面的语句使更改生效

[root@hadoop100 include]# source /etc/profile

检查java是否安装成功。如果能看到版本信息就说明安装成功了。
[root@hadoop100 include]# java -version
java version "1.8.0_121"
Java(TM) SE Runtime Environment (build 1.8.0_121-b13)
Java HotSpot(TM) 64-Bit Server VM (build 25.121-b13, mixed mode)
[root@hadoop100 include]#

2. Hadoop安装与环境配置

 Hadoop的安装也是只需要把hadoop的tar包拷贝到linux，解压，设置环境变量.然后用之前做好的xsync脚本，把更新同步到集群中的其他机器。如果你不知道xcall、xsync怎么写的。可以翻一下之前的文章。这样集群里的所有机器就都设置好了。

[root@hadoop100 include]# tar -zxvf /opt/software/hadoop-2.7.3.tar.gz -C /opt/modules/

[root@hadoop100 include]# vi /etc/profile 继续添加HADOOP_HOME

#JAVA_HOME
export JAVA_HOME=/opt/modules/jdk1.8.0_121
export PATH=$PATH:$JAVA_HOME/bin

#HADOOP_HOME
export HADOOP_HOME=/opt/modules/hadoop-2.7.3
export PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin

[root@hadoop100 include]# source /etc/profile

把更改同步到集群中的其他机器

[root@hadoop100 include]# xsync /etc/profile

[root@hadoop100 include]# xcall source /etc/profile

[root@hadoop100 include]# xsync hadoop-2.7.3/

3. Hadoop分布式配置

然后需要对Hadoop集群环境进行配置。对于集群的资源配置是这样安排的，当然hadoop100显得任务重了一点 :) 

编辑0/opt/modules/hadoop-2.7.3/etc/hadoop/mapred-env.sh、yarn-env.sh、hadoop-env.sh 这几个shell文件中的JAVA_HOME，设置为真实的绝对路径。

export JAVA_HOME=/opt/modules/jdk1.8.0_121

打开编辑 /opt/modules/hadoop-2.7.3/etc/hadoop/core-site.xml, 内容如下

<configuration>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://hadoop100:9000</value>
    </property>

    <property>
        <name>hadoop.tmp.dir</name>
        <value>/opt/modules/hadoop-2.7.3/data/tmp</value>
    </property>
  
</configuration>

编辑/opt/modules/hadoop-2.7.3/etc/hadoop/hdfs-site.xml， 指定让dfs复制5份，因为我这里有5台虚拟机组成的集群。每台机器都担当DataNode的角色。暂时也把secondary name node也放在hadoop100上，其实这里不太好，最好能和主namenode分开在不同机器上。

<configuration>
    <property>
        <name>dfs.replication</name>
        <value>5</value>
    </property>

    <property>
        <name>dfs.namenode.secondary.http-address</name>
        <value>hadoop100:50090</value>
    </property>

    <property>
        <name>dfs.permissions</name>
        <value>false</value>
    </property>
</configuration>

YARN 是hadoop的集中资源管理服务，放在hadoop100上。 编辑/opt/modules/hadoop-2.7.3/etc/hadoop/yarn-site.xml

<configuration>

<!-- Site specific YARN configuration properties -->

    <property>
        <name>yarn.nodemanager.aux-services</name>
        <value>mapreduce_shuffle</value>
    </property>


    <property>
        <name>yarn.resourcemanager.hostname</name>
        <value>hadoop100</value>
    </property>

    <property>
        <name>yarn.log-aggregation-enbale</name>
        <value>true</value>
    </property>

    <property>
        <name>yarn.log-aggregation.retain-seconds</name>
        <value>604800</value>
    </property>
</configuration>

为了让集群能一次启动，编辑slaves文件（/opt/modules/hadoop-2.7.3/etc/hadoop/slaves），把集群中的几台机器都加入到slave文件中，一台占一行。

hadoop100
hadoop101
hadoop102
hadoop103
hadoop104

最后，在hadoop100上全部做完相关配置更改后，把相关的修改同步到集群中的其他机器

xsync hadoop-2.7.3/

在启动Hadoop之前需要format一下hadoop设置。

hdfs namenode -format

然后就可以启动hadoop了。从下面的输出过程可以看到整个集群从100到104的5台机器都已经启动起来了。通过jps可以查看当前进程。

[root@hadoop100 sbin]# ./start-dfs.sh
Starting namenodes on [hadoop100]
hadoop100: starting namenode, logging to /opt/modules/hadoop-2.7.3/logs/hadoop-root-namenode-hadoop100.out
hadoop101: starting datanode, logging to /opt/modules/hadoop-2.7.3/logs/hadoop-root-datanode-hadoop101.out
hadoop102: starting datanode, logging to /opt/modules/hadoop-2.7.3/logs/hadoop-root-datanode-hadoop102.out
hadoop100: starting datanode, logging to /opt/modules/hadoop-2.7.3/logs/hadoop-root-datanode-hadoop100.out
hadoop103: starting datanode, logging to /opt/modules/hadoop-2.7.3/logs/hadoop-root-datanode-hadoop103.out
hadoop104: starting datanode, logging to /opt/modules/hadoop-2.7.3/logs/hadoop-root-datanode-hadoop104.out
Starting secondary namenodes [hadoop100]
hadoop100: starting secondarynamenode, logging to /opt/modules/hadoop-2.7.3/logs/hadoop-root-secondarynamenode-hadoop100.out
[root@hadoop100 sbin]# jps
2945 NameNode
3187 SecondaryNameNode
3047 DataNode
3351 Jps
[root@hadoop100 sbin]# ./start-yarn.sh
starting yarn daemons
starting resourcemanager, logging to /opt/modules/hadoop-2.7.3/logs/yarn-root-resourcemanager-hadoop100.out
hadoop103: starting nodemanager, logging to /opt/modules/hadoop-2.7.3/logs/yarn-root-nodemanager-hadoop103.out
hadoop102: starting nodemanager, logging to /opt/modules/hadoop-2.7.3/logs/yarn-root-nodemanager-hadoop102.out
hadoop104: starting nodemanager, logging to /opt/modules/hadoop-2.7.3/logs/yarn-root-nodemanager-hadoop104.out
hadoop101: starting nodemanager, logging to /opt/modules/hadoop-2.7.3/logs/yarn-root-nodemanager-hadoop101.out
hadoop100: starting nodemanager, logging to /opt/modules/hadoop-2.7.3/logs/yarn-root-nodemanager-hadoop100.out
[root@hadoop100 sbin]# jps
3408 ResourceManager
2945 NameNode
3187 SecondaryNameNode
3669 Jps
3047 DataNode
3519 NodeManager
[root@hadoop100 sbin]#

4. Hadoop的使用

使用hadoop可以通过API调用，这里先看看使用命令调用，确保hadoop环境已经正常运行了。

 这中间有个小插曲，我通过下面的命令查看hdfs上面的文件时，发现连接不上。

[root@hadoop100 ~]# hadoop fs -ls
ls: Call From hadoop100/192.168.56.100 to hadoop100:9000 failed on connection exception: java.net.ConnectException: Connection refused; For more details see: http://wiki.apache.org/hadoop/ConnectionRefused

后来发现，是我中间更改过前面提到的xml配置文件，忘记format了。修改配置后记得要format。

hdfs namenode -format

hdfs 文件操作

[root@hadoop100 sbin]# hadoop fs -ls /
[root@hadoop100 sbin]# hadoop fs -put ~/anaconda-ks.cfg /
[root@hadoop100 sbin]# hadoop fs -ls /
Found 1 items
-rw-r--r--   5 root supergroup       1233 2019-09-16 16:31 /anaconda-ks.cfg
[root@hadoop100 sbin]# hadoop fs -cat /anaconda-ks.cfg

文件内容

[root@hadoop100 ~]# mkdir tmp
[root@hadoop100 ~]# hadoop fs -get /anaconda-ks.cfg ./tmp/
[root@hadoop100 ~]# ll tmp/
total 4
-rw-r--r--. 1 root root 1233 Sep 16 16:34 anaconda-ks.cfg

执行MapReduce程序

hadoop中指向示例的MapReduce程序，wordcount，数数在一个文件中出现的词的次数，我随便找了个anaconda-ks.cfg试了一下：

[root@hadoop100 ~]# hadoop jar /opt/modules/hadoop-2.7.3/share/hadoop/mapreduce/hadoop-mapreduce-examples-2.7.3.jar wordcount /anaconda-ks.cfg ~/tmp
19/09/16 16:43:28 INFO client.RMProxy: Connecting to ResourceManager at hadoop100/192.168.56.100:8032
19/09/16 16:43:29 INFO input.FileInputFormat: Total input paths to process : 1
19/09/16 16:43:29 INFO mapreduce.JobSubmitter: number of splits:1
19/09/16 16:43:30 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1568622576365_0001
19/09/16 16:43:30 INFO impl.YarnClientImpl: Submitted application application_1568622576365_0001
19/09/16 16:43:31 INFO mapreduce.Job: The url to track the job: http://hadoop100:8088/proxy/application_1568622576365_0001/
19/09/16 16:43:31 INFO mapreduce.Job: Running job: job_1568622576365_0001
19/09/16 16:43:49 INFO mapreduce.Job: Job job_1568622576365_0001 running in uber mode : false
19/09/16 16:43:49 INFO mapreduce.Job: map 0% reduce 0%
19/09/16 16:43:58 INFO mapreduce.Job: map 100% reduce 0%
19/09/16 16:44:10 INFO mapreduce.Job: map 100% reduce 100%
19/09/16 16:44:11 INFO mapreduce.Job: Job job_1568622576365_0001 completed successfully
19/09/16 16:44:12 INFO mapreduce.Job: Counters: 49
File System Counters
FILE: Number of bytes read=1470
FILE: Number of bytes written=240535
FILE: Number of read operations=0
FILE: Number of large read operations=0
FILE: Number of write operations=0
HDFS: Number of bytes read=1335
HDFS: Number of bytes written=1129
HDFS: Number of read operations=6
HDFS: Number of large read operations=0
HDFS: Number of write operations=2
Job Counters
Launched map tasks=1
Launched reduce tasks=1
Rack-local map tasks=1
Total time spent by all maps in occupied slots (ms)=6932
Total time spent by all reduces in occupied slots (ms)=7991
Total time spent by all map tasks (ms)=6932
Total time spent by all reduce tasks (ms)=7991
Total vcore-milliseconds taken by all map tasks=6932
Total vcore-milliseconds taken by all reduce tasks=7991
Total megabyte-milliseconds taken by all map tasks=7098368
Total megabyte-milliseconds taken by all reduce tasks=8182784
Map-Reduce Framework
Map input records=46
Map output records=120
Map output bytes=1704
Map output materialized bytes=1470
Input split bytes=102
Combine input records=120
Combine output records=84
Reduce input groups=84
Reduce shuffle bytes=1470
Reduce input records=84
Reduce output records=84
Spilled Records=168
Shuffled Maps =1
Failed Shuffles=0
Merged Map outputs=1
GC time elapsed (ms)=169
CPU time spent (ms)=1440
Physical memory (bytes) snapshot=300003328
Virtual memory (bytes) snapshot=4159303680
Total committed heap usage (bytes)=141471744
Shuffle Errors
BAD_ID=0
CONNECTION=0
IO_ERROR=0
WRONG_LENGTH=0
WRONG_MAP=0
WRONG_REDUCE=0
File Input Format Counters
Bytes Read=1233
File Output Format Counters
Bytes Written=1129
[root@hadoop100 ~]#

 在web端管理界面中可以看到对应的application：

执行的结果，看到就是“#” 出现的最多，出现了12次，这也难怪，里面好多都是注释嘛。

[root@hadoop100 tmp]# hadoop fs -ls /root/tmp
Found 2 items
-rw-r--r--   5 root supergroup          0 2019-09-16 16:44 /root/tmp/_SUCCESS
-rw-r--r--   5 root supergroup       1129 2019-09-16 16:44 /root/tmp/part-r-00000
[root@hadoop100 tmp]# hadoop fs -cat /root/tmp/part-r-0000
cat: `/root/tmp/part-r-0000': No such file or directory
[root@hadoop100 tmp]# hadoop fs -cat /root/tmp/part-r-00000
#    12
#version=DEVEL    1
$6$JBLRSbsT070BPmiq$Of51A9N3Zjn/gZ23mLMlVs8vSEFL6ybkfJ1K1uJLAwumtkt1PaLcko1SSszN87FLlCRZsk143gLSV22Rv0zDr/    1
%addon    1
%anaconda    1
%end    3
%packages    1
--addsupport=zh_CN.UTF-8    1
--boot-drive=sda    1
--bootproto=dhcp    1
--device=enp0s3    1
--disable    1
--disabled="chronyd"    1
--emptyok    1
。。。

通过web 界面可以查看hdfs中的文件列表 http://192.168.56.100:50070/explorer.html#

 hadoop还有好多好玩儿的东西，等待我去发现呢，过几天再来更新。